Coping with uncertainty on health decisions: assessing new solutions

Abstract

Uncertainty is present in most human decisions. The introduction of health technologies in publicly funded systems is no exception. Sources of uncertainty in this domain come from different grounds, and a variety of means and tools have been developed to cope with unknown parameters and uncertain variables. Clinical trials were established more formally back in the 1950s and 1960s of the last century in most developed countries as a requirement to reduce uncertainties in general and primarily to guarantee safety and, later, efficacy. The outcomes of such clinical trials allowed us to understand the features of a given agent under certain controlled specific conditions, usually by comparing its medical performance versus a placebo. They have provided basic knowledge for registration, and to fix the price and reimbursement conditions for the majority of new drugs for several decades. However, in order to use this information on drugs, health authorities and physicians had to assume that the biological response of individuals to pharmaceutical agents and the clinical management of patients were similar across jurisdictions so that health outcomes held generally. That is to say, the basic and implicit assumption was that health outcomes from clinical trials were fully transferable to other health systems where decisions on the utilization of that drug had to be adopted. Needless to say, this basic assumption was subject to strong criticism and raised many questions with regard to whether similar health outcomes in different medical environments could be achieved. As a result of this debate, efficacy outcomes needed to be supplemented with real world data from other studies, i.e., data on effectiveness. In this sense, databases such as the Cochrane library—constituting a world reference since 1988—provide the medical community with real practice data. Additionally, once therapeutic areas became crowded with several similar agents, relative efficacy data started to be demanded by health authorities and physicians so that uncertainties about which drug would be more appropriate for each patients subgroup could be clarified. Although an old concern, this information requirement was considered more intensively during the 1980s and demand continues to grow. Furthermore, most current comparisons of the efficacy and safety of a new treatment with existing ones are still carried out using indirect methods, with the results again being subject to uncertainty (the already cited Cochrane library offers a collection of evidence that is easy to access but still many questions remain open to research). These uncertainties come from a variety of sources, such as differences in recruitment criteria, clinical management of patients in each trial, study duration, the statistical design of the trial, and so on. In this context, comparative efficacy studies (whose outcomes come from face-to-face clinical trials) began to be designed in response to requests from many health bodies to some rather reluctant pharmaceutical companies, who were unwilling to compare their new molecules with other, well-established treatments. This new route towards the reduction of uncertainty is still clearly under development and has a long way to go.